|
| | | Forum Newbie
Group: Forum Members
Last Login: 4/30/2008 3:50:16 PM
Posts: 5, Visits: 5 |
| My SP1 install went well except for two exceptions.
One, my http content active monitor no longer works. I had it set on two of my devices and it worked before I installed SP1. But now both devices show the http content monitor as failing, even though my webservers are running fine. I have even tried removing and reinstalling the monitors and the problem remains.
The second problem is similar to one that another poster already mentioned. I have a maintenance window set on a device (lets call it routerA) for which other devices are up dependent. One of the devices that is up dependent (lets call it serverA) on routerA is sending up alerts when it shouldn't. The maintenance window is from 5:00am to 7:30am every day of the week. This morning I had an alert from serverA at 7:32am showing it was up. My up alerts are set to only trigger if a device has been down for at least 5 minutes. Since serverA is up dependent on routerA which was in a maintenance window until 7:30am, I shouldn't have received any up alert at 7:32am, should I have? I have about 20+ maps set up with this same configuration and this was the only up alert I received, so the problem does not appear to be widespread.
-john
|
| |
| | | Junior Member
Group: Forum Members
Last Login: 3/11/2005 5:43:00 PM
Posts: 14, Visits: 1 |
| Commenting just on your first problem.. the HTTP Content Scan.
In this message I replace "<" with "x" and ">" with "y" so that things will still render ok in the forum.
The (interesting part of the) definition before upgrade was:
Expect=.*x/htmly
This says "I want 0 or more occurences of any character followed by the string x/htmly". Now, "x/htmly" should occur right at the end of a webpage. So the Content Scan Monitor downloads the whole page. May be a small page. May be a huge page.
The WhatsUp Professional SP1 definition changes to:
Expect=xhtmly
Which is "Match against xhtmly anywhere in the data". Even if your server is a bit over the top with its HTTP headers then will normally appear in the first packet or two of response data. And if it does then WUP stops downloading the rest of the data at that point.
I just did a test to compare the two monitors on a sample page (Internet, not local):
1) xhtmly in 1st data packet. Timestamp = 0.15 seconds
2) First rendered character. Timestamp = 0.27 seconds
3) x/htmly. Timestamp = 0.41 seconds
So, The SP1 HTTP Content Scan definition has the ability to be much faster.
However, there are a couple of things to keep in mind:
1) An upgrade to SP1 will not modify the existing monitor definition. The xhtmly incarnation of the monitor is only delivered with a fresh install.
2) Both definitions should be regarded as demonstrations of possibilities. The Knowledgebase explains how to tweak the definition in order to retrieve a specific page. Modify the Expect to look for specific text that is only rendered if the webserver's backend database is working and you have created a monitor that tells you that both webserver and database look healthy.
Me? I like speed. I place a file MONITOR.TXT file on my webserver. A file that contains only the string "Hello World". Then I tailor the monitor to retrieve MONITOR.TXT and match against the string. No overhead at all.
Mark Symons
Ipswitch, Inc
Augusta GA |
| |
|
|